---
language: 
  - en

tags:
- pytorch
- ner
- text generation
- seq2seq

inference: false

license: mit

datasets:
- conll2003

metrics:
- f1
---
# t5-base-qa-ner-conll

Unofficial implementation of [InstructionNER](https://arxiv.org/pdf/2203.03903v1.pdf).
t5-base model tuned on conll2003 dataset.

https://github.com/ovbystrova/InstructionNER 

## Inference 
```shell
git clone https://github.com/ovbystrova/InstructionNER 
cd InstructionNER
```

```python
from instruction_ner.model import Model
model = Model(
    model_path_or_name="olgaduchovny/t5-base-ner-mit-movie",
    tokenizer_path_or_name="olgaduchovny/t5-base-ner-mit-movie"
)
options = [
        "ACTOR",
        "AWARD",
        "CHARACTER",
        "DIRECTOR",
        "GENRE",
        "OPINION",
        "ORIGIN",
        "PLOT",
        "QUOTE",
        "RELATIONSHIP",
        "SOUNDTRACK",
        "YEAR"
    ]
instruction = "please extract entities and their types from the input sentence, " \
              "all entity types are in options"
text = "are there any good romantic comedies out right now"
generation_kwargs = {
    "num_beams": 2,
    "max_length": 128
}
pred_spans = model.predict(
    text=text,
    generation_kwargs=generation_kwargs,
    instruction=instruction,
    options=options
)
>>> [(19, 36, 'GENRE'), (41, 50, 'YEAR')]
```